From AI-Generated Worksheets
to Learner Autonomy

Hello everyone. My name is Yen. Today I share a small project. I used Generative AI to design English worksheets for young learners. The topic is: from AI-generated worksheets to learner autonomy. In 10 slides, I will tell you what I did, what I learned, and what surprised me.

Why this project? Two reasons. First, the challenge: kids need practice at home, but teachers are busy. Second, the opportunity: GenAI can make worksheets very fast — but is the quality good enough? To find out, I worked with 4 Grade-2 students, age 7 to 8, A1 level. I used Action Research. I gave them AI-made worksheets at home, watched them in class, and asked them to fill a short feedback form.

My first lesson. Before this project, I thought good materials simply means fun activities. After this project, I think differently. Fun is not enough. Good materials need four things. One, sequencing — tasks build step by step. Two, scaffolding — give help, then take it away. Three, recycling — students meet the language again and again. Four, balancing — not too easy, not too hard. Fun activities help, but these four things make real learning.

Here is one example. I chose Present Simple, third-person singular — the famous 's' at the end. I designed 5 tasks. Task 1: recognition — circle the correct verb. Task 2: evaluation — tick correct, cross wrong. Task 3: interpretation — match sentences to pictures. Task 4: controlled practice — fill in the blanks. Task 5: error correction — be the teacher and fix the mistake. AI gave me the activities. But the order, the progression — that came from me. Materials are not just a pile of tasks. They are a journey.

What can GenAI do well, and what can it not do? Strengths: it gives many ideas, very fast. It suggests activity types I never thought of. It saves time. But limits are clear. It does not know my four students. It cannot plan a real learning path. It cannot balance easy and hard. And — very often — it uses words above A1 level. The instructions were too long. The tasks were too hard. So every worksheet needed me to adapt it before use.

My view of AI also changed. Before, I saw AI as a creator — it makes the materials. After, I see AI as a draft tool — it gives me a first version. The teacher still has four jobs: evaluate, adapt, sequence, and take responsibility. AI cannot replace pedagogical judgment, classroom knowledge, or the human feeling for our students. AI is a strong partner, but the teacher stays in charge.

Now, what is learner autonomy? Benson, 2011, says it is the slow ability to take responsibility for your own learning. In my project, autonomy does not mean learning all alone. It means: follow instructions by yourself, do tasks with less help, check your own answers, and feel more confident. To support this, I used three ideas. One — intermingle skills so kids stay interested. Two — one task per day, to keep the load light. Three — recycle the grammar across weeks. The plan ran 5 weeks: vocabulary, grammar, reading, speaking — every week.

What did the students say? The instructions were clear. The activities were fun. They learned new words. They want more worksheets like these. Their favourite skill was speaking. The overall mean score was 3.39 out of 4 — quite high for young learners working at home. But the key observation is this: autonomy did not happen because of AI. It happened because the worksheets had clear guidance, self-checking, and a gentle step-by-step progression. The design did the work.

Three things really surprised me. First: the teacher works more, not less. AI cuts drafting time, but adds evaluation time. The workload moves — it does not disappear. Second: the kids loved it. The mean score was 3.39 out of 4, and they asked for more worksheets. Third, and most unexpected: speaking became their favourite activity. This is strange, because speaking is usually the hardest skill for young learners to do alone at home. So I asked myself: why? Three reasons. One, preparation before performance — they reviewed the prompts at home first, so anxiety dropped. Two, teacher scaffolding — in class I gave guidance and follow-up questions, so confidence grew. Three, positive reinforcement — stickers after four days of activities kept them motivated.

To finish. For young learners, autonomy is not a switch — it is built slowly, through four things together: teacher expertise, structured practice, meaningful support, and the right use of AI. The simple equation is: Teacher Design plus AI Support plus Structured Self-Study equals Greater Learner Autonomy. Three findings to remember. One: AI makes materials development faster. Two: AI cannot replace pedagogical judgment. Three: autonomy comes from thoughtful design, not from the AI itself. Thank you very much for listening. I am happy to take your questions.